NicolasL-S

I've been dreaming of something like this for many years. Stata has many strengths. It has a ridiculously simple syntax, much of the data "cleaning" often happens under the radar, it has a nice gui for models and plots, and visualizing datasets is easy and pleasant. This makes it accessible to undergraduate students without a programming background. It also has obvious downsides. It is proprietary, it can be very slow, and using Mata to write one own's estimation procedure is quite painful (it was last time I did). Julia seems like a promising solution to these problems.

Two potential approach:

1. Calling Julia from Stata like the authors suggest (instead of using mata). This could work if the interface isn't too cumbersome. The ability to compile Julia binaries could also allow to do this without running Julia in the background. Of course, one still needs to pay for Stata.

2. Rewriting Stata in Julia using macros. This sounds like a massive undertaking, but could be the best solution in the long run in my opinion (but would sacrifice the gui).

Great work as always, especially the improvement in error messages.

I think Stan Lee once said about programming languages: "With great power there must also come great binary size."

I would be cautious saying Simulink has no Julia equivalent. This NASA engineer switched from Simulink to SciML and obtained a 15,000x speed up while his code went from 1000 lines to 50 lines long: https://www.youtube.com/watch?v=tQpqsmwlfY0

You don't have to. Just use Rcall.

Of course R has been here longer. Eleven years after its creation, R had fewer than 500 packages. Julia was released in 2012 and today has over 7,000 packages.

> IME languages that are designed as general-purpose languages tend to be better than languages that are designed for niche purposes, even within those niches. The things that make a language good require consistent, general thinking.

Julia is a general purpose language, while being the best tool for scientific computing.

I would suggest adding a bunch of them as such:

zig array + zig variable + zig for loop + zig programming language

Also be careful with the break in the data in 2022.

I'm guessing the choice made by Google went like:

Option 1) Hire more staff to do a proper monitoring job on a per-video or per-channel basis. Option 2) Just demonetize every small youtuber who talks about the Ukraine war right now, no matter the guidelines.

1) costs money. 2) some mild reputation impact for us, but who cares about the livelihood of innocent bystanders who trusted our platform and actually respected our guidelines.

Thousands of channels were demonetized at once. Did you (or Google) checked them before passing judgement?

Among those demonetized is Perun who makes neutral hour-long PowerPoint presentations on war logistics with text, graphs and the occasional harmless picture. His content is higher quality than most news outlet by a margin. This is what he wrote today:

"I am extremely disappointed given that this week's video includes an interview with the former commander of the US Army in Europe, Lieutenant General Ben Hodges (Ret.), on the topic of Ukrainian force generation over the winter and into 2023."

You should check out this case study of migrating Simulink to Julia at NASA: https://www.youtube.com/watch?v=tQpqsmwlfY0

From Simulink to Julia, his code went from 1000 lines to 50 lines and ran 15,000x faster.